INDEX

Explanations

commands and instructions related to coding

oai_token-act-pair · gpt-3.5-turbo

phrases indicating preparation or readiness for action

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

soType

-0.67

 tolerated

-0.65

arium

-0.62

existent

-0.62

respected

-0.61

Unknown

-0.59

 Patriot

-0.59

mart

-0.59

lif

-0.59

nai

-0.58

POSITIVE LOGITS

 proceed

1.02

 commence

0.84

 proceeded

0.84

 Prepare

0.84

 prepare

0.79

 DRAG

0.76

 rinse

0.75

ETHOD

0.75

 explore

0.74

 initialize

0.74

Activations Density 0.280%

commands and instructions related to coding

phrases indicating preparation or readiness for action

No Comments

No Known Activations